Introduction - ICRISAToar.icrisat.org/4830/1/Population improvement in sorghum.pdfimprovement) and...

93

Population improvement in sorghum

Belum VS Reddy and A Ashok Kumar International Crops Research Institute for the Semi-Arid Tropics,

Patancheru 502 324, Andhra Pradesh, India

IntroductionSorghum [Sorghum bicolor (L.) Moench] is the fi fth important cereal crop in the world after wheat, rice, maize and barley. Of late, it has emerged as ‘fuel’ crop in addition to its food, feed and fodder utilities. Sorghum is predominantly a self-pollinated crop and development of new varieties is a natural option for crop improvement. However, there is 5 to 15% outcrossing depending upon the nature of genotype and humidity (House 1985), which makes it possible to use population improvement and hybrid development methods to exploit the heterosis. Discovery of genetic male sterility (GMS) (Table 8) and cytoplasmic-nuclear male sterility (CMS) facilitated the application of recurrent selection procedures (population improvement) and hybrid cultivar development methods, respectively, in sorghum improvement programs.

Sorghum breeders have been successful in exploiting available vast genetic variability in sorghum by direct selection among landraces and/or by the use of conventional methods of handling segregating generations derived from carefully and deliberately effected crosses through pedigree and backcross breeding in the development of pure-line varieties and/or hybrid parents. These conventional breeding methods, used as a short-term strategy produce varieties with a relatively narrow genetic base, favor the accumulation of linkage blocks due to rapid fi xation of genes, and limit recombination options because of continuous inbreeding. On the other hand, population improvement methods, besides offering greater opportunities for recombination to break linkages between desired and undesired traits, provides scope for increased utilization of biotic and abiotic stress resistant, but agronomically non-elite source germplasm lines. The population improvement provides long-term breeding strategy to derive diverse and broad genetic-based superior varieties/hybrid parents (Bola Nath 1982). Therefore, a comprehensive crop improvement strategy has to combine both short- and long-term programs for continuous improvement of economic traits.

J139_1208SorghumImprovementMilleniumF.indd 93J139_1208SorghumImprovementMilleniumF.indd 93 05/12/2008 9:14:18 PM05/12/2008 9:14:18 PM

94

What is population improvement?A population is a group of plants sharing a common gene pool. Population improvement includes (1) the development of broad genetic-based gene pools and (2) its improvement through recurrent selection methods. Recurrent selection was fi rst suggested by Jenkins (1940) and named by Hull (1945). Recurrent selection methods are most suitable for the improvement of those traits that are inherited in a quantitative manner and the essential features of these methods are i) the improvement of the mean performance of the population by increasing the frequency of the genes that effect trait/traits under selection, ii) appearance of new combination of genotypes that never existed in the base population and iii) maintenance of genetic variability by recombination of superior genotypes for further and continuous improvement.

The recurrent selection methods require extensive hybridization, which is tedious to follow in sorghum owing to its inbreeding system. However, the discovery of genetic male-sterility (GMS) and the advent of various mating systems and reciprocal recurrent selection methods in exploiting additive (A) and A×A and other epistatic genetic variation (Comstock and Robinson 1952, Eberhart 1972), led many breeders to adopt population improvement methods in sorghum in the 1960s (Maunder 1972; Doggett 1972a). Several sources of genetic male sterility have been reported from both India and USA, and in all cases it was shown that a recessive allele in homozygous condition at a number of loci with alleles designated as ms1, ms2, ms3, ms4, ms5, ms6, ms7 and al confer male sterility (Table 8). Of these, only ms3 and ms7 alleles have been extensively used in population improvement as they are stable across locations and seasons (Reddy and Stenhouse 1994), although al (antherless) allele is also useful (House 1985).

Development of random-mating populationsPopulations can be developed for different purposes: for improving a single trait; for selecting several traits simultaneously; and for generating fertility restorer and non-restorer (maintainer) populations for deriving hybrid parents. Whatever the purpose, the development of a population involves three steps: (a) selection of component parents (b) introgression of a GMS gene, and (c) random mating among parents.


95

Table 8. Genetic male sterility genes, their designated symbols and mechanism of sterility in sorghum. Source: Adapted from Rooney (2000).Gene symbol Mechanism Referencems1 Normal pollen is dominant over aborted or

empty pollen cellsAyyangar and Ponnaiya (1937)

ms2 Normal pollen is dominant over aborted or empty pollen cells

Stephens (1937)

ms3 Normal pollen is dominant over aborted or empty pollen cells

Webster (1965)

ms4 Empty pollen cells Ayyangar (1942)ms5 Aborted pollen Barabas (1962)ms6 Micro-anthers without pollen Barabas (1962)ms7 Empty pollen cells Andrews and Webster (1971)al Anther less stamens Karper and Stephens (1936)

Population improvement methodsOnce the broad genetic-based populations are developed, their genetic potential can be upgraded theoretically for any number of traits of interest through selection. Selection process is cyclic, with or without progeny testing and a recombination phase involving intercrossing among the selects. The process of selection and the intercrossing among the selects constitute one cycle and many such cycles are repeated and hence the whole process is referred to as recurrent selection. Various population improvement/selection methods have been developed and they could be classifi ed based on the type of population (ie, intra-population and inter-population) and on the units of selection. In intra-population improvement, selection is practiced within a specifi c population for its improvement while in inter-population improvement; selection is based on the intercross performance between two populations. The unit of selection could be individual plants, half-sib families, full-sib families and selfed/inbred (S1 and S2) progenies. Selection based on the visual inspection of individual plant is referred to as phenotypic selection since its genetic worth cannot be known unless the trait under selection has high habitability. On the other hand, selection based on families/progenies is known as genotypic selection, as genetic worth of selected plant is assessed based on its progeny performance. Thus, several intra- and inter-population improvement selection methods are recognized and are described below.


96

Intra-population improvement methodsMass selection. Mass selection is the simplest and easiest of all methods and requires the fewest resources and only one generation per cycle. It is effective, particularly if the trait of interest has high heritability. Jan-orn et al. (1976) predicted that mass selection would be effective in improving highly heritable traits like days to fl ower and plant height in sorghum. It is useful if the population is highly heterogeneous. Mass selection procedures in sorghum have been suggested by Doggett (1968). Mass selection involves selection and recombination of selected plants in a population. A population segregating for male-sterility is planted in isolation. Several superior male-sterile plants are selected based on visual assessment and harvested individually. Equal quantity of seeds from selected plants is bulked to constitute the population for next cycle of selection. Thus, it follows that the unit of selection is individual plant.

Doggett (1972a) has described modifi ed mass selection with alternating male-sterile (female) and male-fertile (male) plants selection in successive generations, aimed at enhancing selection response by increased parental control. In one cycle, seed is harvested from only selected male-sterile plants. These seeds are bulked and sown to constitute the population for the next cycle of selection, wherein, male-fertile plants are selected and harvested seed from selected plants is bulked for selection of male-sterile plants. This procedure is continued. Mass selection should be used in the fi rst few cycles of selection after synthesis of a population. This makes populations reasonably uniform for plant height and maturity before using more sophisticated methods of recurrent selection requiring family/progeny evaluation.

Half-sib family/progeny selection. Half-sib selection requires two generations per cycle since it involves progeny testing. Male-sterile plants are tagged at the time of fl owering and are allowed for open-pollination. Each head is harvested and threshed separately. A part of the seeds from each head is sown in yield trial (evaluation phase) and the remaining is saved as remnant seed. The best entries are chosen from the yield trials, and the remnant seed from these entries is bulked to constitute the population for the recombination phase. Along with the replicated yield trial of half-sib families as well as other families (which will be described subsequently) for selection of best families to recombine, a separate nursery is planted simultaneously to identify male-sterile plants. Sib-mated male-sterile heads are harvested and bulked with remnant seed of families selected on the basis of grain yield and other selection criteria as appropriate in a yield trial. This bulk is sown to allow random mating in the next season. Again, male-sterile plants are tagged and harvested individually to form the next cycle of evaluation. This method of selection is known as “Half-sib family selection” because the unit of selection is half-sib families and a breeder has control over only one of the parents,


97

ie, male-sterile plants (females). Recombination is carried out in the off-season and evaluation in the main season. The method was used with low selection intensities to improve backup populations (eg, Downs Bulk, Brown Population, WABC, Bulk Y and RS5DX) at ICRISAT and progress was made for overall agronomic desirability, grain quality, and in increasing uniformity for plant height and maturity.

Full-sib family/progeny selection. Full-sib families can be developed by crossing selected male-fertile plants onto selected male-sterile plants. The full-sib families so generated are evaluated in a yield trial and the remnant seed of the selected families is then bulked and allowed to recombine. Crosses of male-fertile plants with male-sterile plants are then made and the cycle repeats. In this scheme of selection, the unit of selection is full-sib families and the breeder has the control over both the parents unlike in half-sib family selection.

S1 family/progeny selection. It is one of the most effective selection schemes for sorghum (Gardner 1972). S1 family selection requires three generations per cycle. Heads of male-fertile plants are bagged at fl owering to ensure selfi ng, or they can be tagged to ensure that heads from male-fertile plants and not male-sterile plants are harvested at maturity. Selected plants are harvested and threshed separately, each head forming an S1 family. These families are evaluated in yield trials. Remnant seed from the families selected or their sibbed families based on the yield trials is sown, and seed from male-sterile heads are selected to ensure recombination. Seeds from male-sterile heads are then bulked and sown. Male-fertile heads of good plants are identifi ed for testing to begin next cycle. The units of selection and recombination are S1 progenies. The basic concept behind selfed progeny selection is to expose deleterious recessive genes to facilitate their elimination during evaluation and to increase additive genetic variation. Doggett (1972b) reported the fi rst evidence of success of this method and observed, on an average, 25% yield increase per cycle. After one cycle, the improved population produced a higher grain yield than the best varieties. Jan-orn et al. (1976) predicted that S1 family/progeny testing and selection offered the greatest promise for improvement in NP3R, a population developed at the University of Nebraska, whether calculated on a cycle or on an annual basis. S1 testing is very effi cient if three generations can be grown a year. This is possible only with very early maturing populations.

S2 family/progeny selection. In this scheme of selection process, heads of selected male-fertile plants are bagged at fl owering to ensure selfi ng, or they can be tagged to ensure that heads from male-fertile plants and not male-sterile plants are harvested at maturity. The S1 progenies are grown and plants in S1 progenies rows are selected and again selfed. Selected selfed plants are harvested and threshed separately, each head forming an S2 family. These S2 families are evaluated in yield trials and handled exactly in a manner similar to that in S1 progeny selection


98

and thus selfed/inbred progenies constitute units of selection in both the methods. S2 progeny testing is expected to result in maximum gain per cycle and is most suitable when two growing seasons are available per year, thus permitting one cycle every two years. The method has several advantages over others: additive genetic variance is maximized in S2 families; the families are suffi ciently uniform to permit precise evaluation; two generations per year provide suffi cient time between the generations for sending seed to test locations in a range of environments and analyzing the data for the selection of lines for recombination; selection for different traits can be done in various generations ranging from half-sib to S2 according to the nature of their inheritance; and the lines evaluated are more homozygous and it is hence easier to extract pure lines. In addition, continuous selfi ng and evaluation is expected to improve the probability of deriving more vigorous inbred lines. The disadvantage of the scheme lies in the necessity to sib-mate S2 lines to increase the frequency of male-sterile plants for reconstituting the next cycle of the population.

Testcross family/progeny selection. This selection method is a slight deviation from the concept of intra-population improvement in the sense that the targeted population is improved based on the evaluation of testcross progenies generated by crossing several selected plants of targeted population with a broad based tester population. Thus, the targeted population is not only improved for per se performance but also with respect to general combining ability (gca). It essentially involves three steps; (1) self-pollination of male-fertile plants and testcrossing them to a broad based tester, (2) evaluation of testcross progenies in a yield trial and (3) intercrossing of selected plants based on yield trial.

A number of male-fertile plants from the base population are selected and selfed and simultaneously crossed to a broad based tester. The resultant testcross progenies are evaluated in a yield trial to identify promising families, the units of selection. The remnant seed of selected testcross progenies is bulked and sown and allowed for open pollination with male-sterile plants. Seeds from male-sterile heads are then bulked and sown. Male-fertile heads of good plants are identifi ed for testcrossing to begin next cycle. The cyclic nature of selection through testcrosses (with broad based tester) from selfed seed increases the frequency of genes conferring good gca and gradually improves the population per se performance as well. As the population is improved both for per se performance as well as gca, the probability of deriving vigorous lines with good gca is higher. The choice of the type of tester has been extensively debated. While broad based tester improves gca, a narrow genetic-based tester such as an inbred improves specifi c combining ability (sca). The use of low yielding testers, presumably with low frequency of favorable alleles at important loci, have been advocated because they are expected to increase variation in testcross progenies and thus allow better discrimination among the plants to be evaluated in testcrosses (Hallauer and Miranda 1981).


99

Inter-population improvement methodsHalf-sib reciprocal recurrent selection. Comstock et al. (1949) fi rst suggested the use of Reciprocal recurrent selection (RRS), which maximizes the genetic divergence between the two populations for loci with dominance and/or over-dominance effects. The main objective of this method is to develop the two populations simultaneously so that superior inbred lines would be extracted that combine well with each other. In this scheme of selection, each population provides a source material to advance/improve and also serves as a tester for the other population. Individual selected male-fertile plants (tag or bag these plants for identifi cation at maturity) in one population, designated as ‘A’ will be crossed to several random male-sterile plants of the other population designated as ‘B’. In a similar manner, several selected male-fertile plants of population ‘B’ are crossed onto several random male-sterile plants of population ‘A’. The crosses thus generated are evaluated in a yield trial and seeds from selected male-fertile plants are bulked and grown in isolation. Incorporate heterozygous male-sterile plants into these populations. Allow for random pollination of male-sterile heads. Mark male-sterile plants at fl owering. Harvest seed from male-sterile plants in each population and bulk to constitute the new populations from which male-fertile plants would be selected and crossed to male-sterile plants from the other population and the cycle repeats. This method is useful in sorghum since hybrids are commercially viable and large inter-population heterosis is observed. There are two types of RRS depending on the identity of the parents involved in the crosses; (1) half-sib RRS and (2) full-sib RRS. Half-sib RRS is the most promising in sorghum because it provides a better evaluation of males to be selected. In half-sib RRS method, random male-fertile plants (their identity is not maintained) of one population are crossed onto male-sterile plants of the other population to generate half-sib families for evaluation in yield trial and thus half-sib families form units of selection. The subsequent steps are as described above.

Full-sib RRS. In this method, only selected male-fertile plants of one population (their identity is maintained) are crossed onto several selected male-sterile plants of the other population to generate full-sib families for evaluation in yield trial and thus full-sib families become units of selection. Since identity of both female and male plants is known, this method is called full-sib RRS. The use of RRS in sorghum, especially full-sib RRS, is hampered by the sterility system used to enable random mating. All crosses (test and selection units) would be generated using the male-sterile as the female for which no selfed seed can be produced. Thus, from the selected full-sibs, only the male parents from each cross can be used as recombination units, effectively reducing selection intensity and failing to capture genes from those female parents producing superior crosses.

Reciprocal full-sib selection has been used to improve the sorghum populations KP9BC0 and GTPP7R, a derivative of TP24. Nearly 200 reciprocal full-sibs from


100

each population (with the other population as male parent) were tested collaboratively by Kansas State University (KSU) at Garden City and the University of Nuevo Leon in Monterrey, Mexico. The top 15% full-sibs were selected from the best stress site, and remnant S1 seed of the male parents of the selected full-sibs was used for recombination. Estimates of genetic variances, heritabilities, and intra-population predicted gains were reported by Chisi (1993). Estimates for the genetic variability and mean were found to be consistently higher in the TP24×KP9B (TP24 as female) than the KP9B×TP24 reciprocal crosses, suggesting that signifi cant cytoplasmic effects may exist (Rattunde et al. 1997).

Population improvement programs at ICRISATPopulation improvement programs were followed extensively at ICRISAT during the initial periods of its inception. However, during later periods, it received less emphasis due to changes in funding patterns, donors’ emphasis on short-term impact and the requirement of national programs. Population improvement programs at ICRISAT, Patancheru, Andhra Pradesh, India were initiated in 1973 with the introduction of a large number of populations (ICRISAT 1974) from Kansas State, Nebraska and Purdue Universities (USA), Serere (Uganda), Nigeria and from Australia using genetic male-sterility induced by ms3 and ms7 genes following recurrent selection procedures to breed for wider adaptability. New backup and advanced populations were synthesized by intercrossing selected progenies from populations of similar maturity, geographic origin, and restoration behavior (Bhola Nath 1977). Backup populations were selected under low selection intensity to maintain variability for a long time. However, the backup populations were later discontinued to reduce the size of the program. The advanced populations were subjected to rigorous selection, with the objective of producing superior varieties and hybrid parents. Initially when resources were abundant, half-sib/s1/s2 family selection methods were followed to improve the populations. However, due to subsequent diminishing resources, simple mass selection alternated with recombination methods became corner stones of developing trait-specifi c broad genetic-based populations (Reddy et al. 2004). Over the years, an array of 19 populations (Fig. 25) for long-term improvement for key agronomic traits or trait combinations and resistance to major insect pests and diseases were developed using ms3 and ms7 genes and improved at ICRISAT, Patancheru (Reddy et al. 2005).

The populations developed by ICRISAT and NARS collaborators in the Southern African region target contrasting agro-ecological zones (ICRISAT 1989, Rattunde et al. 1997). Guinea and caudatum and guinea×caudatum populations have been developed in Mali (Rattunde et al. 1997). In Southern Africa, the four random mating populations developed jointly by ICRISAT and national agricultural research systems (NARS) using ms7 gene provided broad genetic-based gene pools from


101

which national programs and South African Development Committee (SADC) could develop improved lines and varieties using recurrent selection (Obilana 1989).

Utilization of populationsThe economic benefi ts of population improvement are ultimately realized when readily usable genetic variability for traits of interest from these populations is effectively exploited to develop pure-lines and/or hybrid parents. It is towards this goal that populations are improved by recurrent selection. It is assumed that as the mean performance of a population is improved, there will be a parallel improvement in the performance of its derived progenies (Eberhart 1972). The probability of deriving pure-lines with desired combination of genes is higher from an improved population with higher frequency of desirable alleles. Studies at ICRISAT have given good indications that it will work. A set of random S1 progenies from different cycles of two populations were evaluated for grain yield in two trials (Prasit 1981). The contribution of the most advanced cycles in each population is the highest, followed by the previous cycle, indicating that as the average grain yield of populations increased, the grain yield of the derived lines also increased.

Fig. 25. A high-tillering forage population developed at ICRISAT, Patancheru.


102

Traditional pedigree selection methods used during the inbreeding produce pure-lines for direct use as varieties or hybrid parental lines or, more frequently, as improved parental lines for use in pedigree breeding activities. Purging of GMS gene/s from the derived lines is essential if it/they has/have been used for building the populations. Selection against the male-sterile gene can be easily handled by identifying sterile plants at fl owering. The lines from the populations are produced by successive selfi ng of male-fertile plants at any stage until the progeny becomes uniform. Continued selection for male-fertile plants eliminates male sterility from the lines. This approach has been used in the development of lines from S1 families originating in several populations at Purdue (Rattunde et al. 1997), and in the ongoing derivation of restorer lines and dual-purpose varieties out of the US/R (DP) population at ICRISAT.

At ICRISAT, the process of identifying superior lines began in the early stages of population development. For example, a line from the Diallel population has been released as Melkamash in Ethiopia (Table 9). A foliar resistant line A 2267-2 is derived from US/R population at ICRISAT and is extensively used as one of the parents in sorghum breeding programs in China, Ethiopia and Latin America. Several hundred pairs of cytoplasmic-nuclear male-sterility-based seed parents (A-) and their corresponding maintainer lines (B-) have been developed in an array of maturity and plant types. The male sterile line SPL 132A (renamed as 421A by China) developed from Diallel population at ICRISAT is directly used as female parent in development and release of fi ve hybrids, Liao Za 4, 6, 7 and 10 and Jin Ja 94, in China (Table 9).

Table 9: Examples of successful utilization of populations at ICRISAT. Variety Population Place of release Released name Year of releaseDiallel Pop-7-682 Diallel Ethiopia Melkamash 1982A 3681 FLR China Yuan 1-98 1982A 3872 Bulk Y China Yuan 1-28 1982A 3895 Bulk Y China Yuan 1-505 1982A 6072 Indian Synthetic China Yuan 1-54 1982

SPL 132A (renamed as 421A in China)

Diallel China Liao Za -4, -6, -7 and -10 and Jin Ja 94

1988

M 36248 WABC Myanmar YEZIN 2 (Schwephyu 2) 1984Source: Reddy et al. 2004.


103

ReferencesAndrews DJ and Webster OJ. 1971. A new factor for genetic male-sterility in Sorghum bicolor (L.) Moench. Crop Sci. 11:308–309.

Ayyangar GNR. 1942. The description of crop plant characters and their ranges of variation. IV. Variability of Indian sorghum. Indian J. Agric. Sci.12:527–563.

Ayyangar GNR and Ponnaiya BWX. 1937. The occurrence and inheritance of purple pigment on the glumes of sorghum close on emergence from the boot. Curr. Sci. 5:590.

Barabas Z. 1962. Observation of sex differentiation in sorghum by use of induced male-sterile mutants. Nature. 195:257–259.

Bhola Nath. 1977. Populations breeding in sorghum. Presented at the International Sorghum Workshop, 6−12 March 1977. ICRISAT, Patancheru, India.

Bhola Nath. 1982. Population breeding techniques in sorghum. Pages 421−434 in Sorghum in eighties. Proceedings of the International Symposium on Sorghum, International Crops Research Institute for the Semi-Arid Tropics (ICRISAT), 2−7 November 1981 (House LR, Mughogho LK and Peacock JM, eds.). Patancheru 502 324, Andhra Pradesh, India: International Crops Research Institute for the Semi-Arid Tropics.

Comstock RE, Robinson HF and Harvey PH. 1949. A breeding procedure designed to make maximum use of both general and specifi c combining ability. J. Am. Soc. Agron. 41:360−367.

Comstock RE and Robinson HF. 1952. Genetic parameters, their estimation and signifi cance. Pages 284−291 in Proceedings of the Sixth International Grasslands Congress, 17-23 August 1952, Pennsylvinia State College, Pa., USA.

Doggett H. 1968. Mass selection systems for sorghum. Crop Sci. 8:391−392.

Doggett H. 1972a. The improvement of sorghum in East Africa. Pages 47–59 in Sorghum in Seventies (Rao NGP and House LR, eds.). New Delhi: Oxford and IBH Publishing Co.

Doggett H. 1972b. Recurrent selection in sorghum populations. Heredity 28:9−29.

Eberhart SA. 1972. Techniques and methods of more effi cient population improvement in sorghum. Pages 197−213 in Sorghum in Seventies (Rao NGP and House LR, eds.). New Delhi, India: Oxford and IBH Publishing Co.

Gardner CO. 1972. Development of superior populations of sorghum and their role in breeding programs. Pages 180−196 in Sorghum in Seventies (Rao NGP and House LR, eds.). New Delhi, India: Oxford and IBH Publishing Co.

Hallauer AR and Miranda JB. 1981. Quantitative genetics in maize breeding. Ames, Iowa: Iowa State University Press. 468 pp.

House LR. 1985. A guide to sorghum breeding (2nd edition). Patancheru 502 324, Andhra Pradesh, India: International Crops Research Institute for the Semi-Arid Tropics.

Hull FH. 1945. Recurrent selection for specifi c combining ability in corn. J. Am. Soc. Agron. 37:134−145.


104

ICRISAT. 1974. Pages 14−21 in Annual Report 1973/74. Patancheru 502 324, Andhra Pradesh, India: International Crops Research Institute for the Semi-Arid Tropics.

ICRISAT. 1989. Pages 7−25 in Annual Report 1988. Patancheru 502 324, Andhra Pradesh, India: International Crops Research Institute for the Semi-Arid Tropics.

Jan-orn J, Gardner CO and Ross WM. 1976. Quantitative genetic studies of the NP3R random mating grain sorghum population. Crop Sci. 16:489−496.

Jenkins MT. 1940. The segregation of genes affecting yield of grain in maize. J. Am. Soc. Agron. 32:55−63.

Karper RE and Stephens JC. 1936. Floral abnormalities in sorghum. J. Hered. 27:183–194.

Maunder AB. 1972. Objectives and Approaches to grain and forage sorghum improvement in the Americas. Pages 60–100 in Sorghum in Seventies (Rao NGP and House LR, eds.). New Delhi: Oxford and IBH Publishing Co.

Obilana AB. 1989. Progress in Sorghum Breeding in the SADC Region (1984/85−1987/88) and Plans for 1988/89. Pages 213−248 in Proceedings of the Fifth Annual Workshop on Sorghum and Millets for Southern Africa, 21−23 September 1988, Maseru, Lesotho. PO Box 776, Bulawayo, Zimbabwe: SADC/ICRISAT Sorghum and Millet Improvement Program.

Prasit J. 1981. Effect of recurrent selection in two sorghum populations. MSc thesis, Andhra Pradesh Agricultural University (APAU), Rajendranagar, Hyderabad, Andhra Pradesh, India.

Rattunde HFW, Weltzien RE, Bramel-Cox PJ, Kofoid K, Hash CT, Schipprack W, Stenhouse JW and Presterl T. 1997. Population improvement of pearl millet and sorghum: current research, impact and issues for implementation. In Proceedings of the International Conference on Genetic Improvement of Sorghum and Pearl Millet, 22−27 Sepember 1996, Lubbock, Texas, USA.

Reddy BVS and Stenhouse JW. 1994. Sorghum improvement for the semi-arid tropic region: past, current and future research thrusts in Asia. Punjabrao Krishi Vidyapeeth Res. J. 18:155–170.

Reddy BVS, Rao Prakasha, Deb UK, Stenhouse JW, Ramaiah B and Ortiz R. 2004. Global sorghum genetic enhancement processes at ICRISAT. Pages 65−102 in Sorghum Genetic Enhancement: Research Process, Dissemination and Impacts, Patancheru 502 324, Andhra Pradesh, India: International Crops Research Institute for the Semi-Arid Tropics.

Reddy BVS, Ramesh S and Ortiz R. 2005. Genetic and cytoplasmic-nuclear male-sterility in sorghum. Plant Breeding Reviews 25:139–172.

Rooney WL. 2000. Genetics and cytogenetics. Pages 261–307 in Sorghum: Origin, History, Technology and Production (Smith CW and Frederiksen RA, eds.). New York: Wiley.

Stephens JC. 1937. Male sterility in sorghum: Its possible utilization in production of hybrid seed. J. Am. Soc. Agron. 29:690–696.

Webster OJ. 1965. Genetic studies in Sorghum vulgere (Pers.). Crop Sci. 5:207–210.


Date post:	03-Jul-2020
Category:	Documents
Upload:	others
View:	0 times
Download:	0 times

Introduction - ICRISAToar.icrisat.org/4830/1/Population improvement in sorghum.pdfimprovement) and...

Documents